Adobe Analytics Expands Coworker, MCP Access and Analysis Workspace

In October 2026, Adobe introduced a series of Adobe Analytics updates that expand the use of AI for analyzing user behavior and managing analytical reports. The principal changes concern CX Enterprise Coworker, differentiated access to the MCP server, and improvements to Analysis Workspace.

On October 29, Coworker Chat capabilities for analyzing Adobe Analytics data with natural-language questions, investigating changes in metrics, and transferring visualizations into Analysis Workspace became generally available. On October 6, Adobe introduced a separate MCP Read-only Access permission, letting external AI assistants run analytical queries without permission to modify Adobe Analytics components. Adobe also updated Workspace visualizations, journey comparisons, component descriptions, and classification processing in Data Warehouse.

CX Enterprise Coworker: Analyze Adobe Analytics With Natural Language

On October 2, Adobe began rolling out the integration of CX Enterprise Coworker Chat with Adobe Analytics. The capability became generally available on October 29. Coworker provides conversational access to Report Suite data. Instead of manually creating tables and segments, users can ask an analytical question in natural language—for example, “Which traffic sources generated the largest increase in revenue over the past 30 days?” Coworker queries the available Adobe Analytics data and returns an answer with relevant metrics.

Supported questions include work that would previously have required direct use of Analysis Workspace, such as investigating conversion trends, comparing periods, exploring audience segments, and assessing acquisition channels. This can reduce the manual work involved in routine analysis, especially for people who are less familiar with Adobe Analytics. Coworker does not create missing data. It works with events, metrics, and dimensions already collected in the relevant Report Suites. If interactions were never tracked, or tracking was incorrect, the AI interface cannot fix the underlying data gap on its own.

Root Cause Analysis: Investigating Changes in Metrics

Adobe began rolling out Root Cause Analysis for Adobe Analytics in Coworker Chat on October 13; general availability followed on October 29. The tool helps users move beyond describing a metric change to exploring the factors associated with it.

For example, an analyst might ask, “Why did revenue spike on Tuesday?” Coworker examines Report Suite data and identifies dimensions or segments related to the observed change, potentially revealing which traffic sources, audience categories, or other available attributes contributed most. A decline in registrations can similarly be broken down by device, acquisition channel, or page to narrow the investigation.

Root Cause Analysis identifies statistical factors associated with a change; it does not necessarily establish causation. An increase in mobile conversions does not prove that device type caused the change. Confirming the explanation can require additional evidence, campaign-change reviews, or A/B testing.

Open Coworker Visualizations in Analysis Workspace

Another feature began rolling out on October 13 and became generally available on October 29: visualizations built by Coworker Chat from Adobe Analytics data can now be opened in a new Analysis Workspace project.

A marketer could ask Coworker to chart orders by advertising channel and then move that visualization into Workspace for further exploration, with additional segments, different levels of detail, and full reporting capabilities. Coworker thus serves as a starting point, while the conventional analytical tools remain available for verification and deeper work. For teams, a marketer can identify a pattern conversationally and an analyst can continue the investigation in Workspace. The generated visualization is not automatically a validated finding; metric definitions, date ranges, and conversion calculations still need review.

Migrate Adobe Analytics Projects to Customer Journey Analytics With Coworker

On October 26, Adobe began rolling out Coworker-assisted migration of Adobe Analytics projects to Customer Journey Analytics (CJA), with general availability on October 29. Previously, migration was carried out through the Adobe Analytics interface or Adobe Analytics 2.0 API.

Coworker checks whether dimensions and metrics in the source project align with the target CJA Data View, helps identify field mappings where needed, and then recreates the project in CJA. The process covers related components as well as the project structure. Organizations with many Workspace reports on traffic, registrations, and revenue can reduce some of the manual work required when transitioning to CJA.

Users with the Component Migration permission can initiate migration through Coworker even if they are not Adobe Analytics administrators. Migrating a project does not guarantee identical numbers across the two platforms. Adobe Analytics and CJA may use different data models, metric definitions, and event-processing rules. Results must be validated after migration.

MCP Read-only Access: Restrict AI Clients to Analytical Queries

On October 6, Adobe introduced a more granular permissions model for the Adobe Analytics MCP Server. MCP lets compatible AI clients query Adobe Analytics tools and perform supported operations. Previously, these capabilities were governed by a single MCP Access permission. Adobe divided access into two levels:

MCP Read-only Access: search for components, read data, run reports, and adjust session parameters.
MCP Full Access: all read operations, plus the creation and modification of supported analytical components.

The former MCP Access permission was renamed MCP Full Access, and existing Full Access users retained their previous capabilities. Read-only is designed for scenarios in which an AI assistant should inspect statistics without modifying the analytics configuration. An analyst with Read-only Access can ask a compatible MCP client to compare conversion rates by channel, but cannot use write-restricted tools to create a segment, change a calculated metric, or save a Workspace project.

MCP Permissions Still Depend on Adobe Analytics Product Rights

MCP permissions do not replace standard Adobe Analytics authorization. A user still needs the relevant product-level rights to access Report Suites and perform specific operations. Read-only does not grant universal access to every Report Suite, while Full Access does not override underlying restrictions on component creation or modification.

Permissions from multiple Product Profiles are additive. If any assigned profile grants MCP Full Access, the user receives the Full Access level. Administrators therefore need to review all assignments, not just a single profile. This matters particularly for agencies and larger organizations connecting external AI tools to client analytics.

Read-only reduces the risk of unintentional configuration changes, but it does not eliminate the risk of data exposure: the connected assistant can still read data the user is authorized to view. Organizations must also assess AI-client selection, data-processing policies, and Report Suite access.

Adobe Analytics MCP Server: Other October and November Changes

Adobe also updated MCP server capabilities throughout the period:

October 1: localization parameters became available in selected component-search and reporting tools, allowing results to reflect the user's locale.
October 5: MCP Access was renamed MCP Full Access and MCP Read-only Access was added, according to the MCP changelog. The Adobe Analytics release notes identify October 6 for the permissions release.
October 19: an optional compressOutput parameter let compatible tools return shorter responses when requested.
November 2: findSegments and findCalculatedMetrics gained filterByName for name-based searches; listComponentUsage gained limit, page, startTime, and endTime for pagination and time-bounded usage analysis.

Adobe also improved segment creation and validation, made setSessionDefaults available with MCP Read-only Access, and enhanced behavior when the MCP server response cache is temporarily unavailable. Session defaults do not require permission to modify analytical objects. These changes support more controlled and efficient AI-assisted analysis.

Journey Canvas: Compare Conversion Rates in Percentage Points

On October 19, Adobe introduced Compare Percentages in the Journey Canvas visualization in Analysis Workspace. The setting represents changes between the current and comparison periods in percentage points.

For example, if 40% of visitors advanced from a product page to the next step in the previous period and 45% did so in the current period, the increase is five percentage points. The corresponding relative percentage increase is 12.5%. These are different measures, and confusing them can distort funnel reporting.

Compare Percentages applies to nodes, connections, and Fallout. It helps teams examine shifts in transition rates rather than relying only on absolute visitor counts, especially when traffic volumes differ across periods. Interpretation still depends on date ranges, audience definitions, and the events included in the journey.

Concise Legend Labels: Simplify Complex Charts

On October 29, Adobe added Concise Labels to Analysis Workspace visualizations. The setting removes repeated elements from legend labels and keeps the distinguishing parts, making charts with numerous series or comparison groups easier to read.

For example, a chart displaying Revenue across multiple advertising channels may otherwise repeat the same metric, segment, and period details in every legend entry. Concise Labels supports Area, Bar, Combo, Donut, Histogram, Horizontal Bar, Line, and Venn charts, including stacked variants. The option is off by default and changes only the presentation of labels, not the source data or metric calculations.

AI-Generated Descriptions for Analytical Components

On October 29, Adobe began rolling out automatic component descriptions in Analysis Workspace, with general availability planned for November 18, 2026. The feature covers Dimensions, Metrics, Calculated Metrics, Segments, and Date Ranges. Users can generate descriptions individually or in batches, helping teams document large collections of custom components.

Adobe also supports generated descriptions for Workspace projects when saving them or through Project Info & Settings. This can simplify maintenance and knowledge transfer. However, AI-generated descriptions need review: an assistant may misunderstand a calculated metric whose true meaning depends on organization-specific business rules. The feature does not replace a shared metric dictionary.

Data Warehouse: More Accurate Classification Handling

On October 14, Adobe began rolling out Data Warehouse improvements for Classification Data, with full rollout planned for January 2027. The changes address two behaviors: deleted classification values should return blanks rather than persisting in later reports, and reports should be able to include all classification values associated with an item rather than showing only one in some cases.

Adobe warns that these changes may produce differences between Data Warehouse and Analysis Workspace. In certain cases, Data Warehouse can show multiple values while Analysis Workspace displays a single value. These differences matter for organizations using classifications to report on advertising campaigns, product categories, and other business dimensions. A change in reported classifications does not necessarily mean user behavior changed; it may reflect corrected processing logic.

Do Not Confuse Adobe Analytics and Customer Journey Analytics Features

Adobe is developing AI capabilities for both Adobe Analytics and Customer Journey Analytics, but they do not have identical functionality. On October 29, Adobe introduced an embedded Coworker panel within Analysis Workspace for CJA. This panel can use an open project as context and modify its tables, panels, segments, visualizations, and date ranges through conversation.

According to the documentation, the embedded CJA panel remained in Limited Testing, was disabled by default, and required activation by a CJA administrator. The general availability of Coworker Chat for Adobe Analytics on October 29 does not mean this same embedded project-editing interface was available to all classic Adobe Analytics customers. Similarly, Conversation Insights for AI-chat interactions belongs to CJA and should not automatically be attributed to Adobe Analytics.

Availability of the October Updates

October 6, 2026: MCP Read-only Access.
October 19, 2026: Journey Canvas Compare Percentages.
October 29, 2026: Coworker analysis of Adobe Analytics, Root Cause Analysis, opening visualizations in Workspace, project migration to CJA, and Concise Legend Labels.
November 18, 2026 (planned): general availability of AI-generated component descriptions.
January 2027 (planned): completion of the Data Warehouse classification rollout.

Adobe Analytics follows a continuous-delivery model, so account-level access can vary by organization and enabled features. Coworker also requires appropriate permissions and access to CX Enterprise Coworker. MCP connections require Adobe Analytics permissions and a compatible AI client.

Adobe Analytics: AI Analysis With Enterprise Reporting Controls

Adobe Analytics is a digital analytics platform for measuring visitor behavior, acquisition sources, conversions, and other business metrics. The October releases advance three areas: Coworker makes existing analytical data more accessible through natural-language questions and investigation of changing metrics; MCP Read-only Access permits connections to external AI tools without granting configuration changes; and Workspace updates improve journey comparisons, chart readability, and component documentation.

Ecommerce teams can use these capabilities to explore purchases, revenue, and conversion behavior; service businesses can investigate lead sources and acquisition performance; B2B teams can analyze product-content engagement, sign-ups, and other conversion events. AI functionality cannot substitute for correct event collection, data governance, and validation of analytical hypotheses. The overall benefit is a shorter path from asking a question to obtaining data and continuing the investigation, while retaining existing permissions and analytical tools.

Trust Us Transform Your Performance Marketing

Turn digital analytics into reliable decisions across acquisition, conversion, and revenue. We help teams implement event tracking, define consistent metrics, connect marketing data sources, and investigate performance changes so dashboards, AI-assisted analysis, and attribution support measurable business goals.

Discuss Your Analytics Strategy